Production condor in the cloud
Download
1 / 26

Production Condor in the Cloud - PowerPoint PPT Presentation


  • 95 Views
  • Updated On :

Production Condor in the Cloud. Ian D. Alderman CycleComputing http://twitter.com/cyclecomputing. Case Studies. This is a presentation of some work we’ve recently completed with customers. Focus on two particular customer workflows: Part 1: 10,000 core cluster “Tanuki.”

Related searches for Production Condor in the Cloud

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about 'Production Condor in the Cloud' - miach


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Production condor in the cloud l.jpg

Production Condor in the Cloud

Ian D. Alderman

CycleComputing

http://twitter.com/cyclecomputing


Case studies l.jpg
Case Studies

  • This is a presentation of some work we’ve recently completed with customers.

  • Focus on two particular customer workflows:

    • Part 1: 10,000 core cluster “Tanuki.”

    • Part 2: Easily running the same workflow “locally” and in the cloud including data synchronization.

  • We will discuss the technical challenges.


Case 1 10 000 cores tanuki l.jpg
Case 1: 10,000 Cores “Tanuki”

  • Run time = 8 hours

  • 1.14 compute-years of computing executed every hour

  • Cluster Time = 80,000 hours = 9.1 compute years.

  • Total run time cost = ~$8,500

  • 1250 c1.xlarge ec2 instances ( 8 cores / 7-GB RAM )

  • 10,000 cores, 8.75 TB RAM, 2 PB of disk space

  • Weighs in at number 75 of Top 500 SuperComputing list

  • Cost to run = ~ $1,060 / hour


Customer goals l.jpg
Customer Goals

  • Genentech: “Examine how proteins bind to each other in research that may lead to medical treatments.”

    - www.networkworld.com

  • Customer wants to test the scalability of CycleCloud: “Can we run 10,000 jobs at once?”

  • Same workflow would take weeks or months on existing internal infrastructure.


Customer goals cont l.jpg
Customer Goals (cont)

  • They can’t get answers to their scientific questions quickly enough on their internal cluster.

    • Need to know “What do we do next?”

  • Genentech wanted to find out what Cycle can do:

    • How much compute time can we get simultaneously?

    • How much will it cost?


Run timeline l.jpg
Run Timeline

  • 12:35 – 10,000 Jobs submitted and requests for batches cores are initiated

  • 12:45 – 2,000 cores acquired

  • 1:18 – 10,000 cores acquired

  • 9:15 – Cluster shut down


System components l.jpg
System Components

  • Condor (& Workflow)

  • Chef

  • CycleCloud custom CentOS AMIs

  • CycleCloud.com

  • AWS


Technical challenges aws l.jpg
Technical Challenges: AWS

  • Rare problems:

    • If a particular problem occurs .4% of the time, if you run 1254 instances it will happen 5 times on average.

    • Rare problem examples:

      • DOA instances (unreachable).

      • Disk problems: can’t mount “ephemeral” disk.

  • Need chunking: “Thundering Herd” both for us and AWS.

  • Almost certainly will fill up availability zones.


Technical challenges chef l.jpg
Technical Challenges: Chef

  • Chef works by periodic client pulls.

    • If all the pulls occur at the same time, run out of resources.

    • That’s exactly what happens in our infrastructure when you start 1250 machines at once.

    • Machines do ‘fast converge’ initially.

    • So we need to use beefy servers, configure appropriately, and then stagger launches at a rate the server can handle.

  • This was the bottleneck in the system

    • Future work: pre-stage/cache chef results locally to reduce initial impact on the server.


Technical challenges cyclecloud l.jpg
Technical Challenges: CycleCloud

  • Implemented monitoring and detection of classes of rare problems.

  • Batching of requests with delays between successive requests.

  • Testing: better to request 256 every 5 minutes or 128 every 2.5 minutes? What’s the best rate to make requests?

  • Set chunk size to 256 ec2 instances at a time

    • Did not overwhelm AWS/CycleCloud/Chef infrastructure

    • 2048 cores got job work stream running immediately

    • 1250 ec2 requests launched in 25 minutes ( 12:35am – 12:56 am Eastern Time )


  • Technical challenges images fs l.jpg
    Technical Challenges: Images + FS

    • Images not affected by scale of this run.

    • Filesystem: large NFS server.

    • We didn’t have problems with this run (we were worried!)

    • Future work: parallel filesystem in the cloud.



    Technical challenges condor13 l.jpg
    Technical Challenges: Condor

    • Basic configuration changes.

      • See condor-wiki.cs.wisc.edu

    • Make sure user job log isn’t on NFS

      • Really a workflow problem…

    • Used 3 scheduler instances (could have used one)

      • 1 main scheduler

      • 2 auxilary schedulers

      • Worked very well and handled the queue of 10,000 jobs just fine

        • Scheduler instance 1: 3333 jobs

        • Scheduler instance 2: 3333 jobs

        • Scheduler instance 3: 3334 jobs

      • Very nice, steady and even stream of condor job distribution ( see graph )


    So how did it go l.jpg
    So how did it go?

    • Set up cluster using CycleCloud.com web interface.

    • All the customer had to do was submit jobs, go to bed, and check results in the morning.





    Case 2 synchronized cloud overflow l.jpg
    Case 2: Synchronized Cloud Overflow

    • Insurance companies often have periodic (quarterly) compute demand spikes. Want results ASAP but also don’t want to pay for hardware to sit idle the other 11+ weeks of the quarter.

    • Replicate internal filesystem to cloud.

    • Run jobs (on Windows).

    • Replicate results.

    /Users/ianalderman/Old_Desktop/oldDesktop/top/Condor Week/Condor Week 2010-2.ppt


    Customer goals20 l.jpg
    Customer Goals

    • Step outside the fixed cluster size/speed trade off.

    • Replicate internal Condor/Windows pipeline using shared filesystem.

    • Make it look the same to the user – minor UI change to run jobs in Cloud vs. local.

    • Security policy constraints – only outgoing connections to cloud.

    • Continue to monitor job progress using CycleServer.


    System components21 l.jpg
    System components

    • Condor (& Workflow)

    • CycleServer

    • Chef

    • CycleCloud custom Windows AMIs / Filesystem


    Technical challenges images and file system l.jpg
    Technical Challenges: Images and File System

    • Customized Windows images (Condor).

    • Init scripts a particular problem.

    • Machines reboot to get their hostname.

    • Configure SAMBA on Linux Condor scheduler.


    Technical challenges chef23 l.jpg
    Technical Challenges: Chef

    • Implemented support for a number of features for Chef on Windows that were missing.

    • Lots of special case recipes because of the differences.

    • First restart caused problems.


    Technical challenges condor24 l.jpg
    Technical Challenges: Condor

    • DAGMan submission has three stages.

      • Ensure input gets transferred.

      • Submit actual work (1000-5000 compute-hours)

      • Transfer results

    • Cross platform submission.


    Technical challenges cycleserver l.jpg
    Technical Challenges: CycleServer

    • CycleServer is used:

      • As a job submission point for cloud jobs.

      • To monitor the progress of jobs.

      • To monitor other components (Collect-L/Ganglia).

      • To coordinate file synchronization on both ends.

        • Custom plugins.

        • DAG jobs initiating file synchronization – wait until it completes.

        • Plugins do push/pull from local to remote due to firewall.


    Cloud vs on demand clusters l.jpg
    Cloud vs. On-demand Clusters

    Cloud Cluster

    Physical Cluster

    Actions taken to provision: Button pushed on website

    Duration to start: 45 minutes

    Copying software – minutes

    Copying Data – minutes to hours

    Ready for Jobs

    Actions Taken to provision

    • Budgeting

    • Eval Vendors

    • Picking hardware options

    • Procurement process (4-5 mo)

    • Waiting for Servers to ship

    • Getting Datacenter space

    • Upgrading Power/Cooling

    • Unpacking, Stacking, Racking the servers

    • Plugging networking equipment and storage

    • Installing images/software

    • Testing Burn-in and networking addrs

    • Ready for jobs


    ad